Matrices are used in large-scale statistical computing systems and have many applications. In several applications, matrices are used to represent multi-dimensional datasets. Many such data sets include large numbers of data observations but still larger numbers of missing data observations or data points that are otherwise represented by zero, including features of a data observation that are zero. Such matrices are referred to as sparse matrices.